ABSTRACT In a seminal paper, Sarwal et al.1 reported that the presence of a B cell gene signature and dense clusters of B cells in the renal biopsies was strongly associated with glucocorticoid-resistant T cell-mediated graft rejection (TCMR) and kidney allograft loss. They proposed a hypothesis that, despite standard of care immunosuppression, infiltrating B cells play a pivotal role in a subset of acute cellular rejection of kidney allografts in the clinic. While some subsequent studies supported this hypothesis, others reported opposite outcomes. As a result, the role of intrarenal B cells in clinical TCMR is currently unclear, and an understanding of their role is urgently required to allow clinicians to rationally decide if B cell depletion might be useful for reversing steroid-resistant TCMR. We hypothesize that the identification of additional features of intrarenal B cells or of the rejecting biopsy are required to more accurately predict poor graft outcome and the need for B cell depletion to treat TCMR. These features include an understanding the specificity of intrarenal B cells and their functional properties, as well as the inflammation architecture of the biopsy. To this end, we have developed two innovative approaches to understand in situ adaptive T and B cell immunity in human biopsies from rejecting kidney allografts. First, we are able to pair the transcriptional state in single B cells with the antigenic specificity of the antibody they secrete. Sorted B cells from renal biopsies are subjected to single cell (sc) RNA-Seq, and the immunoglobulin variable regions are cloned from these same cells and expressed corresponding antibodies. Second, we have developed a novel approach to characterize the spatial architecture between different immune cell populations in human tissue and identify functional relationships2,3. Using a novel deep convolutional neural network (DCNN), coupled to a tuned neural network (TNN), we can accurately capture T cell shape change upon recognition of antigen presented by dendritic cells (DCs) in multicolor confocal micrographs. Therefore, this analytic pipeline (Cell Distance Mapping, CDM) can identify and quantify antigen presentation in fixed tissue samples from both mice and humans. We will use CDM to study T cell interactions with B cells or dendritic cells (DCs) in rejecting kidney biopsies. Third, we developed a mouse model of kidney transplantation to address mechanistic questions raised by observations made from the human renal biopsies. We will apply these three approaches to test our project hypothesis in two Aims: Aim 1. Test the hypothesis that the different transcriptional and functional states of intrarenal B cells during rejection are associated with B cell receptor (BCR) specificity and rejection type. Aim 2: Test the hypothesis that the type of cellular rejection (T cell mediated rejection vs. mixed T cell- and antibody-mediated rejection) is defined by characteristic in situ cellular interactions between T cells and B cells or DCs that are amenable to specific therapeutic intervention.